A while back (couple years) I started tracking how much coffee I drank, and (roughly) when.

I created a simple Mojolicious webapp, putting data on a SQLite database. I later expanded it to accept/use an "API key" with which I could track my coffee intake, or get how much I drank already and when was my last cup, "simply" by running a simple bash+curl script. The site, instead, displays a neat github-commits-like chart[1] and some recent stats. I have a branch which changes the code to work on Postgres and is deployable on dokku, but I've not yet made the switch to that. I'm still using a mojolicious+sqlite+hypnotoad process running on a cheap VPS.

Since I started tracking my intake, it looks like I drank, on average, 3-4 coffees per day:

    sqlite> select (select count(*) from intake)/(select julianday(max(time_t)) - julianday(min(time_t)) from intake);
    3.7864677548175

Here in Italy's northern region, the quarantine started region-wide on the 8th of March.

    sqlite> select (select count(*) from intake where time_t > "2020-03-08")/(select julianday(max(time_t)) - julianday("2020-03-08") from intake);
    4.20309898028925

It looks like I'm drinking a bit more coffee than the usual, but surely nowhere as near as my most intake:

    sqlite> select date(time_t) as ymd, count(*) as cnt from intake group by ymd order by cnt desc limit 5;
    2018-05-16|11
    2018-03-12|10
    2018-05-15|10
    2018-02-07|9
    2019-06-04|9

Hope it helps! :)

[1]: https://i.imgur.com/c0DqAGO.png